Mailing List Archives
Authenticated access
|
|
|
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Condor-users] Getting certain jobs to avoid eachother
- Date: Mon, 12 Dec 2011 11:20:08 -0700
- From: Erik Erlandson <eje@xxxxxxxxxx>
- Subject: Re: [Condor-users] Getting certain jobs to avoid eachother
On Mon, 2011-12-12 at 17:49 +0000, Martin Steele wrote:
> I have a class of jobs that I'd like to ensure is maximally dispersed
> amongst my pool's resources.
A memory-specific approach using dynamic provisioning would be to
declare partitionable slots using 'memory=100%'
http://research.cs.wisc.edu/condor/manual/v7.6/3_13Setting_Up.html#sec:SMP-dynamicprovisioning
And then in your job submission files include:
request_memory=<memory-requirement-estimate>
A more generic "job/cpu" oriented approach would be to use partitionable
slots with 'cpu=100%'
And then in your job submission file:
rank=Cpus
Which will heuristically prefer (partitionable) slots with the fewest
cpus already used.